If you want to use this to extract data you will have to make a developer account for Spotify API, then you can get your CLient Secret and Credential to be used for Pulling requests, also note that the Acess is for temporary time limit, then you will have to run the above code again to fetch access
We have taken the list of artist doing good in the charts of last.FM, then we get their data on SpotifyWe will also tap into Billboard charts to see if our artists are there too ?bThen we fill look out for them in the similar playlists in Spotify
We will combine the Dataframes for our modeling purpose and analyze the data by comparing the audio features and popularity for both the tracks.
So we found out that there are 986 dupliactes tracks which is possible as the tracks may be reuploaded with the same name but if they have same track_id that means they are just duplicate
I believe Key is a very Interesting Features as we can see which Chords or Keys are Mostly Popular or tend to be less popular. We can see that the tracks in last.fm list tend to be high in popularity, with Key 5(E) & 9(G#/ A$)having mostly popular songs. Also, tracks in Major keys are more in percentage than minor keys
We will map the audio features in a radar graph to compare songs_lastfm and song_11k
For now we don't identify any outliers in this data
The difference between popularity levels is getting clearer. As we can see that the Basic Probablity for getting a song with more than 40 popularity is around 50% more for Last_fm Artists in comparison of dtat of random artist's tracks